The aim of this work is to provide the cultural heritage community with a comprehensive hyperspectral image database of handwritten laboratory samples, including various writing inks commonly found in historical documents. The database contains 195 samples registered in the VNIR (400-1000 nm) and SWIR (900-1700 nm) spectral ranges, along with complete information about the ink recipes (components and concentrations used for each ink and mixture), and their corresponding Ground Truth images. The database is now publicly available as part of a bigger database related to the Hyperdoc project and can be used to perform different tasks. We present here one example: the classification of iron gall vs non-iron gall inks.
The purpose of this work is to present a new dataset of hyperspectral images of historical documents consisting of 66 historical family tree samples from the 16th and 17th centuries in two spectral ranges: VNIR (400-1000 nm) and SWIR (900-1700 nm). In addition, we performed an evaluation of different binarization algorithms, both using a single spectral band and generating false RGB images from the hyperspectral cube.
Beyond RGB is a free, opensource, software application providing colorimetric and spectral processing of a 6-channel spectral image. The software has an input of two sets of RAW RGB images, one set for each of two different lighting conditions. These sets include a dark current, flatfield, target, and item. The outputs are an RGB image that is color calibrated with data on the accuracy of the calibration and user-selected spectral reflectance estimations of regions of interest. The improvements created for this version of the software include an updated user interface, auto-sorting of files, improved color difference calculation and visualization, a userfriendly website, and the inclusion of various RAW file types.
Three endmember extraction methods (NFINDR, NMF and manual extraction) are compared in two stages (pre- and post- intervention) of the same painting, a Maternity on copper plate, under study for the formulation of a hypothesis on the authorship and the dating. The endmembers are extracted from spectral images in the 400-1000 nm range. The main aim is to determine if simple automatic endmember extraction is enough for pigment and re-painted areas identification in this case study.
Transparent glass frames are often used to exhibit, handle, and store ancient manuscripts (folia or fragments) across museums, libraries, and collections. Once the manuscripts are carefully sealed (glazed), the process of re-opening the frame for the analysis of the glazed manuscript is not always desirable, given their fragile state of preservation. Therefore, microimaging with IR and UV light sources above the glass frame is a frequently used method for the preliminary (qualitative) classification of the inks applied on the manuscripts. Building on this well-established methodology, this study explores the potential of spectral imaging technology for the quantitative analysis of glazed manuscripts. The present research focuses on the colorimetric analysis of iron-gall and carbon black inks applied on a papyrus substrate, aiming to the quantitative analysis of the effect of glass frames to the acquired images. The obtained results show that the quantitative colorimetric analysis of the inks above the glass frame can be used for the preliminary classification of the inks, hence minimizing the need to open the glass frames for further analysis.
A software application for colorimetric and spectral processing of six-channel spectral images has recently been developed. The application, called Beyond RGB, takes as input two RAW RGB image sets (object/flat-field/dark current) captured under two different lighting or filtering conditions, and outputs 1.) a color managed RGB image with ancillary information about the accuracy of the color calibration and 2.) a spectral reflectance transform that enables the interactive estimation of reflectance curves from user-selected regions of interest in the image. Beyond RGB was designed with considerations for form, function, and user friendliness, and is intended for use by cultural heritage imaging professionals. It is cross-platform compatible and is operated through an interactive graphical user interface. Beyond RGB is a living, updatable, open-source project, and is freely available for download from the project’s public GitHub repository.
An imaging process is described which captures spectral reflectance for reflective media. The ultimate target media are prints and photographs within the collection of the Library of Congress. The system is based on a fifteen channel LED source and a monochrome camera. The LED source sequentially illuminated reference and verification targets, with an image captured for each LED channel. From the measured data and images of reference targets, a model was developed to predict spectral reflectance. With that model, the 15 images of a test sample were combined to a single 31-band spectral image. Spectral images can be used to calculate colorimetric data for each pixel, or to better understand material properties. The colorimetric results show that the system predicts good color as compared to the most relevant FADGI guidelines.
Decreasing the use of pesticides is one of the main goals of current agriculture, which requires fast, precise and continuous assessments of crop pests. Citrus pests cause a lot of damage worldwide and the techniques to evaluate them are mainly based on manual, time-consuming readings of insects stuck on traps spread over the crops. This is the case of red scale insects, whose control is notably challenging due to their small size and high reproduction rate. Hence, in this work, we carry out a spectral characterization of this insect in the visible range through spectrometric devices, microscopy and hyperspectral imaging technology to analyze the feasibility of using this information as a means of automatically identifying specimens belonging to this species in this era of precision agriculture. The results obtained show that spectral reflectance differences between red scales and other insects can be recorded at long (red) wavelengths and that red scales are morphologically different, i.e., smaller and more rounded. A reflectance ratio computed from spectral images taken at 774 nm and 410 nm is proposed as a new approach for automated discrimination of red scales from other insects.
This article considers the joint demosaicing of colour and polarisation image content captured with a Colour and Polarisation Filter Array imaging system. The Linear Minimum Mean Square Error algorithm is applied to this case, and its performance is compared to the state-of-theart Edge-Aware Residual Interpolation algorithm. Results show that the LMMSE demosaicing method gives statistically higher scores on the largest tested database, in term of peak signal-to-noise ratio relatively to a CPFA-dedicated algorithm.
Pigment classification of paintings is considered an important task in the field of cultural heritage. It helps to analyze the object and to know its historical value. This information is also essential for curators and conservators. Hyperspectral imaging technology has been used for pigment characterization for many years and has potential in its scientific analysis. Despite its advantages, there are several challenges linked with hyperspectral image acquisition. The quality of such acquired hyperspectral data can be influenced by different parameters such as focus, signal-to-noise ratio, illumination geometry, etc. Among several, we investigated the effect of four key parameters, namely focus distance, signal-to-noise ratio, integration time, and illumination geometry on pigment classification accuracy for a mockup using hyperspectral imaging in visible and near-infrared regions. The results obtained exemplify that the classification accuracy is influenced by the variation in these parameters. Focus distance and illumination angle have a significant effect on the classification accuracy compared to signal-to-noise ratio and integration time.